JP2013195713A - Speech correction device, speech correction method, and computer program for speech correction - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a speech correction device capable of correcting muffled sound in spite of a small difference between a power spectrum in a low frequency band and a power spectrum in a high frequency band.SOLUTION: A speech correction device 6 includes: an effective spectrum extraction unit 13 which, with respect to at least a first frequency band and a second frequency band out of a plurality of frequency bands, calculates effective spectrum signals representing perceivable spectrum signal values by subtracting masking thresholds corresponding to spectrum signal values which cannot be heard in accordance with human auditory characteristics, from spectrum signal values; an inter-band power difference calculation unit 14 which obtains a difference between the effective spectrum signal of the first frequency band and the effective spectrum signal of the second frequency band; a correction amount calculation unit 15 which determines a correction amount for a prescribed frequency band in accordance with the difference; and a correction unit 16 which corrects spectrum signal values at respective frequencies within the prescribed frequency band in accordance with the correction amount.

Description

  The present invention relates to a sound correction device, a sound correction method, and a sound correction computer program for correcting a sound signal, for example.

  In an audio signal collected by a mobile phone, for example, a high frequency component may be relatively small due to a frequency characteristic of a microphone included in the mobile phone. In such a case, when the collected audio signal is reproduced, the reproduced sound becomes a so-called muffled sound, and as a result, it may be difficult for the listener to hear the reproduced sound.

  In order to solve the above-described problems, a technique for enhancing speech without degrading speech quality has been studied (for example, see Patent Document 1).

  For example, the speech enhancement device disclosed in Patent Document 1 calculates SNR, which is a component ratio between received speech and ambient noise, and determines the speech from at least one of the pitch frequency of the received speech and the slope of the power spectrum of the speech. Calculate brightness. In addition, this speech enhancement device masks the received speech from ambient noise from the band division information and SNR indicating the bandwidth that contributes to the improvement of subjective intelligibility of the received speech and the bandwidth that contributes to the improvement of subjective brightness. The enhancement amount of the first band that contributes to the improvement of the subjective intelligibility of the received voice at the time is calculated. Furthermore, the speech enhancement apparatus calculates the enhancement amount of the second band that contributes to the improvement of subjective brightness from the enhancement amount of the first band and the brightness of the speech. The speech enhancement apparatus processes the spectrum of the received speech using the enhancement amount of the first band and the enhancement amount of the second band.

JP 2010-14914 A

In the technique disclosed in Patent Document 1, since the amount of enhancement is affected by the inclination of the power spectrum between the low frequency band and the high frequency band, if the inclination is large to some extent, the listener feels that the sound is muffled. The spectral components in the high frequency band are amplified to the extent that they are not. However, depending on the audio signal, the listener may feel that the sound is congested even if the slope of the power spectrum between the low frequency band and the high frequency band is small. In such a case, the amount of enhancement of the audio signal with respect to the high frequency band is not sufficiently large, and as a result, the listener may feel that the sound is congested even with the enhanced audio signal.
In addition, if the amount of enhancement with respect to the power spectrum inclination is increased, the spectrum component in the high frequency band is excessively amplified for an audio signal having a large power spectrum inclination, and the audio signal is distorted so that it is difficult to hear. .

  In view of this, an object of the present specification is to provide an audio correction apparatus that can correct a muffled sound even if a difference between a power spectrum in a low frequency band and a power spectrum in a high frequency band is small.

  According to one embodiment, an audio correction device is provided. This audio correction apparatus includes a time-frequency conversion unit that calculates a spectrum signal including spectrum signal values for each of a plurality of frequencies by converting a time-domain audio signal into a frequency domain in units of frames having a predetermined time length. A masking threshold value calculation unit that calculates a masking threshold value corresponding to a spectrum signal value that cannot be heard according to human auditory characteristics for each frequency, and at least a frequency included in the first frequency band and a second frequency band An effective spectrum extraction unit for calculating an effective spectrum signal representing a perceptible spectrum signal in the first frequency band and the second frequency band based on the spectrum signal value and the masking threshold for the frequency; Between the bands for obtaining the difference between the effective spectrum signal and the effective spectrum signal of the second frequency band A word difference calculation unit, a correction amount calculation unit that determines a correction amount of a predetermined frequency band according to the difference, and a correction of a spectrum signal value of each frequency within the predetermined frequency band according to the correction amount The correction unit that calculates the correction spectrum signal in step S1 and the frequency time conversion unit that obtains the corrected audio signal by converting the correction spectrum signal into the time domain.

The objects and advantages of the invention will be realized and attained by means of the elements and combinations particularly pointed out in the appended claims.
It should be understood that both the foregoing general description and the following detailed description are exemplary and explanatory and are not restrictive of the invention as claimed.

  The sound correcting device disclosed in the present specification can correct a sound that remains even if the difference between the power spectrum in the low frequency band and the power spectrum in the high frequency band is small.

It is a schematic block diagram of the mobile telephone by which the audio | voice correction apparatus by one Embodiment was mounted. It is a schematic block diagram of an audio | voice correction apparatus. It is a figure which shows an example of the relationship between the peak frequency of a power spectrum, and the masking threshold value of each frequency. (A) is a figure which shows the power spectrum of the input audio | voice signal, and the masking threshold value of each frequency, (b) is the effective power spectrum calculated from the power spectrum of the audio | voice signal shown to (a). It is a figure which shows an example. It is a figure which shows an example of the relationship between the power spectrum of a low frequency band, the power spectrum of a high frequency band, and a power difference. It is a figure which shows an example of the relationship between a power difference and a reference | standard correction coefficient. It is a figure which shows the relationship between a frequency and the coefficient (beta) (f) shown by (5) Formula. (A) is a figure which shows an example of the power spectrum of the audio | voice signal with a feeling of being crowded. (B) is a figure which shows an effective power spectrum among the power spectra shown by (a). (C) is a figure which shows an example of the power spectrum of the audio | voice signal which each corrected the power spectrum shown by (a) with the prior art and the audio | voice correction apparatus by this embodiment. It is an operation | movement flowchart of an audio | voice correction process. It is a figure which shows an example of the relationship between a power difference and a reference | standard correction coefficient by a modification. It is a figure which shows the relationship between the frequency and the coefficient (beta) 1 (f) shown to (7) Formula, and (beta) 2 (f) by the modification. It is a block diagram of the computer which operate | moves as a sound correction apparatus, when the computer program which implement | achieves the function of each part of the sound correction apparatus by embodiment or its modification is operated.

Hereinafter, an audio correction device according to an embodiment will be described with reference to the drawings.
The inventor believes that the difference between the components that can be perceived by humans in the power spectrum of the low frequency band and the high frequency band is more sensitive to the sound than the difference between the power spectrum of the low frequency band and the high frequency band. I got the knowledge that it affects me.
In view of this, the sound correction apparatus obtains a masking threshold corresponding to a power spectrum value that cannot be perceived by a person for each frequency from the power spectrum of each of a plurality of frequencies of the input sound signal. And this audio | voice correction apparatus calculates the effective power spectrum value showing the power spectrum component which a person can perceive by subtracting a masking threshold value from a power spectrum value about each frequency. And this audio | voice correction apparatus makes the amplification factor of the frequency signal value of each frequency contained in a high frequency band high, so that the difference which subtracted the effective power spectrum of a high frequency band from the effective power spectrum of a low frequency band is large.

In the present specification, the term “low frequency band” represents a frequency band that can be perceived by humans for the sake of convenience, which contributes to the deterioration of the feeling of bulkiness of audio signals by increasing the frequency component included in the band. Used for. In addition, the term “high frequency band” is, for convenience, relatively higher than the “low frequency band” and contributes to the improvement of the feeling of volume of the audio signal by increasing the frequency component included in the band. Are used to represent perceptible frequency bands.
Also, the term “power spectrum value” is used to represent the value of the power spectrum for any one frequency. On the other hand, the term “power spectrum” is used to denote a signal sequence of a power spectrum over the entire frequency band including the plurality of frequencies, including the power spectrum value of each of the plurality of frequencies.

FIG. 1 is a schematic configuration diagram of a mobile phone in which the sound correction apparatus according to the first embodiment is mounted. As shown in FIG. 1, a mobile phone 1 includes a control unit 2, a communication unit 3, a microphone 4, an analog / digital converter 5, an audio correction device 6, a digital / analog converter 7, and a speaker. 8.
Among these, the control part 2, the communication part 3, and the audio | voice correction apparatus 6 are each formed as a separate circuit. Alternatively, these units may be mounted on the mobile phone 1 as a single integrated circuit in which circuits corresponding to the units are integrated. Furthermore, each of these units may be a functional module realized by a computer program executed on a processor included in the mobile phone 1.

  The control unit 2 includes at least one processor, a nonvolatile memory, a volatile memory, and its peripheral circuits. When a call is started by an operation via an operation unit (not shown) such as a keypad included in the mobile phone 1, the control unit 2, between the mobile phone 1 and a base station device (not shown), Call control processing such as wireless connection and disconnection is executed according to a communication standard to which the mobile phone 1 complies. Then, the control unit 2 instructs the communication unit 3 to start or end the voice call according to the result of the call control process. Furthermore, the control unit 2 extracts an encoded audio signal included in the downlink signal received from the base station apparatus via the communication unit 3, and decodes the audio signal. Then, the control unit 2 outputs the decoded audio signal to the audio correction device 6 as a received audio signal.

  The control unit 2 encodes a voice signal collected by the microphone 4 and input via the analog / digital converter 5 and generates an uplink signal including the encoded voice signal. Then, the control unit 2 passes the uplink signal to the communication unit 3. In addition, as an encoding method for a speech signal, for example, Adaptive Multi-Rate-NarrowBand (AMR-NB) method standardized by the Third Generation Partnership Project (3GPP), or Adaptive Multi-Rate-WideBand (AMR-WB) method Etc. are used.

  The communication unit 3 performs wireless communication with the base station device. And the communication part 3 receives a radio signal from a base station apparatus, and converts the radio signal into a downlink signal having a baseband frequency. The communication unit 3 performs reception processing such as separation, demodulation, and error correction decoding on the downlink signal, and then passes the downlink signal to the control unit 2. The communication unit 3 performs transmission processing such as error correction coding, modulation, and multiplexing on the uplink signal received from the control unit 2, and then superimposes the uplink signal on a carrier wave having a radio frequency. Transmit to the base station device.

  The microphone 4 is an example of an audio input unit, collects sounds around the mobile phone 1, and generates an analog audio signal corresponding to the intensity of the sound. The microphone 4 outputs the analog audio signal to the analog / digital converter 5.

The analog / digital converter 5 generates a digitized input audio signal by sampling the analog audio signal received from the microphone 4 at a predetermined sampling pitch. Further, the analog / digital converter 5 may include an amplifier, and may be digitized after amplifying the analog audio signal.
The analog / digital converter 5 outputs the input audio signal to the control unit 2.

  The sound correction device 6 calculates the corrected sound signal by enhancing the frequency component in the high frequency band included in the received sound signal so as to suppress the reproduction sound of the received sound signal from being accumulated. Then, the sound correction device 6 outputs the corrected sound signal to the digital / analog converter 7. Details of the sound correction device 6 will be described later.

The digital / analog converter 7 converts the corrected audio signal received from the audio correction device 6 into an analog signal by performing digital-analog conversion. Note that the digital / analog converter 7 may include an amplifier, and the analog corrected sound signal may be amplified by the amplifier. Then, the digital / analog converter 7 outputs the analog corrected audio signal to the speaker 8.
The speaker 8 is an example of an audio output unit, and reproduces the corrected audio signal received from the digital / analog converter 7.

Hereinafter, details of the sound correction device 6 will be described.
FIG. 2 is a schematic configuration diagram of the audio correction device 6 according to one embodiment. The audio correction device 6 includes a time-frequency conversion unit 11, a masking threshold calculation unit 12, an effective power spectrum extraction unit 13, an inter-band power difference calculation unit 14, a correction amount calculation unit 15, a correction unit 16, and a frequency. And a time conversion unit 17.
Each of these units included in the audio correction device 6 may be mounted on the audio correction device 6 as a separate circuit, or may be a single integrated circuit that realizes the functions of these units.

ここで、S(f)は、周波数fにおける周波数信号値であり、F(f)は、周波数fにおけるパワースペクトル値である。なお、周波数信号値及びパワースペクトル値は、それぞれ、スペクトル信号値の一例である。
時間周波数変換部１１は、フレームごとに、各周波数のパワースペクトル値をマスキング閾値算出部１２及び有効パワースペクトル抽出部１３へ出力する。また、時間周波数変換部１１は、各周波数の周波数信号値を補正部１６へ出力する。 The time frequency conversion unit 11 calculates a frequency signal by converting the received audio signal into a frequency domain in units of frames having a predetermined time length (for example, several tens of milliseconds). The frequency signal includes frequency signal values for each of a plurality of frequencies. For this purpose, the time-frequency conversion unit 11 performs time-frequency conversion such as Fast Fourier Transform (FFT) or Modified Discrete Cosine Transform (MDCT) on the received audio signal. To convert it to a frequency signal. Alternatively, the time frequency conversion unit 11 may use a Quadrature Mirror Filter (QMF) filter bank or a wavelet transform as the above time frequency conversion.
The time-frequency converter 11 calculates the power spectrum value of each frequency according to the following equation.

Here, S (f) is a frequency signal value at the frequency f, and F (f) is a power spectrum value at the frequency f. The frequency signal value and the power spectrum value are examples of spectrum signal values.
The time frequency conversion unit 11 outputs the power spectrum value of each frequency to the masking threshold value calculation unit 12 and the effective power spectrum extraction unit 13 for each frame. Further, the time frequency conversion unit 11 outputs the frequency signal value of each frequency to the correction unit 16.

ここで、f_iは、最も低いピーク周波数から順にi番目のピーク周波数を表す。F(f_i)は、ピーク周波数f_iにおけるパワースペクトル値である。そしてm(f,f_i)は、ピーク周波数f_iに基づいて算出される、周波数fのマスキング閾値である。関数α(x)は、変数xが大きくなるほど出力値が単調減少する単調減少関数である。さらに関数max(m(f,f_i))は、各ピーク周波数について算出される周波数fに対するマスキング閾値のうちの最大値を出力する関数である。そしてM(f)は、周波数fのマスキング閾値である。（１）式から明らかなように、マスキング閾値M(f)は、何れかのピーク周波数f_iについて算出されたマスキング閾値m(f,f_i)のうちの最大値となる。 For each frame, the masking threshold value calculation unit 12 calculates a masking threshold value corresponding to a power spectrum value that cannot be perceived by a person based on a person's auditory characteristics for each frequency.
In general, when the spectral component of a sound at a certain frequency is large, the closer to that frequency and the larger the spectral component of that frequency, the more masking effect that makes it difficult to perceive the spectral component of the sound near that frequency. It has been known. This frequency masking effect is attributed to human psychoacoustic characteristics.
Therefore, the masking threshold value calculation unit 12 detects the peak frequency at which the power spectrum value becomes the maximum value by examining the change in the power spectrum value between adjacent frequencies for each frame. Then, the masking threshold value calculation unit 12 calculates the masking threshold value of each frequency according to, for example, the following formula so that the masking threshold value increases as the power spectrum value at the peak frequency increases and the closer to the peak frequency.

Here, f _i represents the i-th peak frequency in order from the lowest peak frequency. F (f _i ) is a power spectrum value at the peak frequency f _i . M (f, f _i ) is a masking threshold value of the frequency f calculated based on the peak frequency f _i . The function α (x) is a monotonically decreasing function in which the output value monotonously decreases as the variable x increases. Furthermore, the function max (m (f, f _i )) is a function that outputs the maximum value of the masking thresholds for the frequency f calculated for each peak frequency. M (f) is a masking threshold of frequency f. (1) As apparent from the equation, the masking threshold M (f) is one of the peak frequency f _i is calculated for the masking threshold m (f, f _i) becomes the maximum value of the.

  FIG. 3 is a diagram illustrating an example of the relationship between the peak frequency of the power spectrum value and the masking threshold value of each frequency set according to the present embodiment. In FIG. 3, the horizontal axis represents frequency and the vertical axis represents power. A graph 301 represents an example of a power spectrum value for each frequency. A graph 302 represents a masking threshold for each frequency. In this example, the power spectrum 301 has local maximum values at the frequencies fA, fB, and fC. That is, the frequencies fA, fB, and fC are peak frequencies. Therefore, as shown in the graph 302, the masking threshold is set to be larger as the frequency is closer to any peak frequency, and to be larger as the power spectrum value at the peak frequency is larger.

The masking effect is also caused by a change in sound volume over time. For example, if the sound in a certain frame is loud, it is difficult to perceive the small sound in the immediately following frame.
Therefore, as a modification, the masking threshold calculation unit 12 may calculate a masking threshold for each frequency based on a change in sound over time. In this case, the masking threshold value calculation unit 12 sets the masking threshold value for each frequency of the current frame, which is the latest frame, to a predetermined number before the current frame, for example, as the power spectrum value of the corresponding frequency in the previous frame increases. Set it to a large value. For example, the masking threshold value calculation unit 12 calculates the threshold value (corresponding to the masking threshold value) described in the C.1.4 Steps in Threshold Calculation section of C.1 Psychoacoustic Model of Annex C of ISO / IEC 13818-7: 2006. Thus, the masking threshold can be calculated. Alternatively, the masking threshold calculation unit 12 may calculate the masking threshold according to the method described in the section of Third Generation Partnership Project (3GPP) TS 26.403 V9.0.0 5.4.2 Threshold Calculation.

According to still another modification, the masking threshold value calculation unit 12 combines, for each frequency, a masking threshold value calculated based on the peak frequency and a masking threshold value calculated based on a change in sound over time. Thus, the final masking threshold may be determined. For example, for each frequency, the masking threshold value calculation unit 12 selects the larger masking threshold value among the masking threshold value calculated based on the peak frequency and the masking threshold value calculated based on the change in sound over time. It may be a masking threshold for.
The masking threshold value calculation unit 12 outputs the masking threshold value for each frequency to the effective power spectrum extraction unit 13.

ここでF(f)は、周波数fのパワースペクトル値であり、M(f)は周波数fのマスキング閾値である。そしてF'(f)は、周波数fの有効パワースペクトル値である。 The effective power spectrum extraction unit 13 is an example of an effective spectrum extraction unit, and for each frame, an effective power spectrum value representing a power spectrum component that can be perceived by a person by subtracting a masking threshold value from the power spectrum value for each frequency. Is calculated. For example, the effective power spectrum extraction unit 13 calculates an effective power spectrum value according to the following equation.

Here, F (f) is a power spectrum value of the frequency f, and M (f) is a masking threshold value of the frequency f. F ′ (f) is an effective power spectrum value of the frequency f.

  FIG. 4A is a diagram illustrating an example of a power spectrum value and a masking threshold value of an audio signal of each frequency. FIG. 4B is a diagram illustrating an example of an effective power spectrum calculated from the power spectrum of the audio signal illustrated in FIG. 4 (a) and 4 (b), the horizontal axis represents frequency, and the vertical axis represents power. A graph 401 represents the magnitude of the power spectrum value of the audio signal at each frequency, and a graph 402 represents a masking threshold value at each frequency. A graph 411 represents the magnitude of the effective power spectrum value of each frequency. The size of the effective power spectrum value represented by the graph 411 is equivalent to the size of the hatched area in FIG.

  The effective power spectrum extraction unit 13 outputs the effective power spectrum value of each frequency to the interband power difference calculation unit 14.

ls及びleは、低周波数帯域の下限及び上限の周波数であり、例えば、ls及びleは、それぞれ、ls以上かつle以下の周波数帯域のパワースペクトルを増幅するとこもり感が悪化する周波数の下限値及び上限値に設定される。一方、hs及びheは、それぞれ、高周波数帯域の下限及び上限の周波数であり、例えば、hs及びheは、それぞれ、hs以上かつhe以下の周波数帯域のパワースペクトルを増幅するとこもり感が改善する周波数の下限値及び上限値に設定される。例えば、ls=150[Hz]、le=800[Hz]、hs=2900[Hz]、he=4000[Hz]に設定される。なお、帯域間パワー差算出部１４は、低周波数帯域及び高周波数帯域内の各周波数の有効パワースペクトル値の中央値を、それぞれ、低周波数帯域及び高周波数帯域の有効パワースペクトルとして、パワー差ΔPを算出してもよい。 The inter-band power difference calculation unit 14 has an effective power spectrum in a relatively low frequency band among frequency bands perceivable by humans and an effective power in a relatively high frequency band among frequency bands perceivable by humans. The power difference from the spectrum is calculated. As the power difference is larger, the high-frequency component that contributes to human perception in the received audio signal is relatively small. Therefore, the power difference is an index representing the degree of muffled received audio signal.
For example, the inter-band power difference calculation unit 14 sets the average value of the power spectrum value of each frequency in the low frequency band and the high frequency band as the effective power spectrum of the low frequency band and the high frequency band, respectively, according to the following equation. Thus, the power difference ΔP is calculated.

ls and le are the lower limit and upper limit frequencies of the low frequency band, for example, ls and le are the lower limit value and the lower limit value of the frequency at which the feeling of obscuration deteriorates when the power spectrum in the frequency band above and below ls is amplified, respectively. Set to the upper limit. On the other hand, hs and he are the lower limit and upper limit frequencies of the high frequency band, respectively.For example, hs and he are frequencies at which the feeling of lumps improves when the power spectrum in the frequency band of hs and higher and he or lower is amplified, respectively. The lower limit value and the upper limit value are set. For example, ls = 150 [Hz], le = 800 [Hz], hs = 2900 [Hz], and he = 4000 [Hz]. Note that the inter-band power difference calculation unit 14 uses the median of the effective power spectrum values of the respective frequencies in the low frequency band and the high frequency band as the effective power spectra of the low frequency band and the high frequency band, respectively, as a power difference ΔP. May be calculated.

  FIG. 5 is a diagram illustrating an example of a relationship between a power spectrum in a low frequency band and a power spectrum in a high frequency band and a power difference. In FIG. 5, the horizontal axis represents frequency and the vertical axis represents power. The graph 500 represents the magnitude of the effective power spectrum value for each frequency. As shown in FIG. 5, the power difference ΔP is a value obtained by subtracting the average value 502 of the effective power spectrum values between the frequency hs and the frequency he from the average value 501 of the effective power spectrum values between the frequencies ls and le. . Therefore, the larger the power difference ΔP, the smaller the ratio of the power spectrum in the relatively high frequency band to the power spectrum in the relatively low frequency band.

  The inter-band power difference calculation unit 14 outputs the power difference ΔP to the correction amount calculation unit 15.

  The correction amount calculation unit 15 emphasizes the frequency signal value of each frequency in the correction target frequency band so as to increase the enhancement degree of the high frequency band with respect to the low frequency band according to the power difference ΔP for each frame. Is determined. In the present embodiment, the correction target frequency band is set so as to include a high frequency band and not include a low frequency band. In the present embodiment, the lower limit frequency of the correction target frequency band is set to a value obtained by subtracting a predetermined offset value from hs, for example, 2562 Hz. In addition, the upper limit of the correction target frequency band is not set.

  First, the correction amount calculation unit 15 obtains a reference correction coefficient Gm that increases as the power difference ΔP increases. For example, the correction amount calculation unit 15 refers to a relational expression or table representing a relationship between the power difference ΔP and the reference correction coefficient Gm, which is stored in advance in a nonvolatile memory circuit included in the correction amount calculation unit 15. A reference correction coefficient Gm is determined.

FIG. 6 is a diagram illustrating an example of the relationship between the power difference ΔP and the reference correction coefficient Gm. In FIG. 6, the horizontal axis represents the power difference ΔP, and the vertical axis represents the reference correction coefficient Gm. A graph 600 represents the relationship between the power difference ΔP and the reference correction coefficient Gm.
As shown in the graph 600, for example, if the power difference ΔP is equal to or less than the reference value Pl, the reference correction coefficient Gm is set to zero. If the power difference ΔP is larger than the reference value Pl and not more than the correction upper limit value Pu, the reference correction coefficient Gm increases linearly as the power difference ΔP increases. When the power difference ΔP becomes equal to or greater than the correction upper limit value Pu, the reference correction coefficient Gm is set to the upper limit value Gmax. Note that the reference value Pl is set to, for example, an upper limit value of a power difference, for example, 28 dB, that does not cause a person to feel a feeling of being bulky with respect to the audio signal without correcting the audio signal. On the other hand, the upper limit value Pu of the correction upper limit Pu and the upper limit value Gmax of the reference correction coefficient are the lower limit value of the power difference at which the distortion of the audio signal caused by enhancing the frequency signal value of each frequency within the correction target frequency band is not subjectively detected, and The upper limit value of the reference correction coefficient is set to 48 dB and 20 dB, for example.

ここで、hsは、上記の高周波数帯域の下限周波数であり、例えば、2900[Hz]に設定される。Esは、補正対象周波数帯域の下限周波数であり、例えば、2500〜2700[Hz]に設定される。またEulは、補正係数g(f)が一定となる周波数の下限値であり、例えば、3100〜3300[Hz]に設定される。 When determining the reference correction coefficient Gm, the correction amount calculation unit 15 multiplies the reference correction coefficient Gm by a coefficient β (f) determined according to the frequency, as shown in the following equation, to thereby calculate each correction frequency band. A correction coefficient g (f) for frequency is determined.

Here, hs is the lower limit frequency of the high frequency band, and is set to 2900 [Hz], for example. Es is a lower limit frequency of the correction target frequency band, and is set to, for example, 2500 to 2700 [Hz]. Eul is a lower limit value of the frequency at which the correction coefficient g (f) becomes constant, and is set to 3100 to 3300 [Hz], for example.

  FIG. 7 is a diagram showing the relationship between the frequency and the coefficient β (f) shown in the equation (5). In FIG. 7, the horizontal axis represents the frequency, and the vertical axis represents the magnitude of the coefficient β (f). The graph 700 represents the relationship between the frequency and the coefficient β (f). As shown in the graph 700, the coefficient β (f) is 0 below the frequency Es, and monotonously increases as the frequency increases above the frequency Es and below the frequency Eul. Β (f) becomes constant when the frequency is greater than the frequency Eul. Thus, by setting the coefficient β (f), in the vicinity of the lower limit Es of the correction target frequency band, the correction coefficient g (f) gradually decreases as the lower limit Es is approached. Therefore, the frequency signal is prevented from becoming discontinuous near the lower limit of the correction target frequency band, so that the corrected sound signal is prevented from being unnaturally distorted.

  The correction amount calculation unit 15 outputs the correction coefficient g (f) of each frequency within the correction target frequency band to the correction unit 16.

S(f)は、周波数fの周波数信号値であり、g(f)は、周波数fの補正係数である。そしてS_out(f)は、補正後の周波数信号値である。（６）式から明らかなように、補正係数g(f)=0のとき、補正後の周波数信号値S'(f)は、補正前の周波数信号値S(f)と等しく、補正係数g(f)が大きくなるほど、補正後の周波数信号値S'(f)は増幅される。 The correction unit 16 corrects the frequency signal value of each frequency within the correction target frequency band in accordance with the following equation in units of frames.

S (f) is a frequency signal value of frequency f, and g (f) is a correction coefficient of frequency f. S _out (f) is the corrected frequency signal value. As apparent from the equation (6), when the correction coefficient g (f) = 0, the frequency signal value S ′ (f) after correction is equal to the frequency signal value S (f) before correction, and the correction coefficient g As (f) increases, the corrected frequency signal value S ′ (f) is amplified.

FIG. 8A is a diagram illustrating an example of a power spectrum of an audio signal having a feeling of being crowded. FIG. 8B is a diagram illustrating an effective power spectrum among the power spectra illustrated in FIG. FIG. 8C is a diagram showing an example of the power spectrum of the audio signal obtained by correcting the power spectrum shown in FIG. 8A with the conventional technology and the audio correcting apparatus according to the present embodiment.
8A to 8C, the horizontal axis represents frequency, and the vertical axis represents power. The graphs 800 shown in FIGS. 8A and 8C represent the power spectrum of an audio signal having a feeling of stagnation for each frequency. Line 801 represents the average value P _low of the power spectrum values of each frequency in the low frequency band, while line 802 represents the average value P _high of the power spectrum values of each frequency in the high frequency band. A graph 810 shown in FIG. 8B represents an effective power spectrum value among the power spectrum values at each frequency shown in FIG. Lines 811 and 812 represent the average value P ′ _low of the effective power spectrum value of each frequency in the low frequency band and the average value P ′ _high of the effective power spectrum value of each frequency in the high frequency band, respectively. A graph 820 shown in FIG. 8C represents a power spectrum value for each frequency of the audio signal corrected according to the conventional technique, and a graph 821 represents the frequency for each frequency of the audio signal corrected by the audio correction device 6. Represents a power spectrum value. As shown in FIG. 8A and FIG. 8B, the effective power spectrum value between the average values P ′ _low and P ′ _high rather than the difference Δ between the average values P _low and P _{high of the} power spectrum values. The difference Δ ′ is larger. Therefore, as shown in FIG. 8C, the audio signal corrected by the audio correction device 6 has a higher frequency band for the power spectrum in the lower frequency band than the audio signal corrected in accordance with the prior art. The power spectrum ratio is increasing. For this reason, the sound signal corrected by the sound correction device 6 has a feeling of being more muffled than the sound signal corrected according to the prior art.

  The correction unit 16 outputs the frequency signal values of all frequency bands including the frequency signal values of each frequency within the corrected correction target frequency band to the frequency time conversion unit 17.

  The frequency-time conversion unit 17 obtains a corrected audio signal by converting the corrected frequency signal value of each frequency into the time domain using the inverse of the time-frequency conversion used by the time-frequency conversion unit 11. . Then, the frequency time conversion unit 17 outputs the corrected audio signal to the digital / analog converter 7.

  FIG. 9 is an operation flowchart of a sound correction process executed by the sound correction device 6. The sound correction device 6 executes sound correction processing for each frame according to the operation flowchart shown below.

  The time-frequency conversion unit 11 calculates the frequency signal by converting the audio signal into the frequency domain in units of frames (step S101). And the time frequency conversion part 11 calculates the power spectrum value of each frequency (step S102). The time frequency conversion unit 11 outputs the power spectrum value of each frequency to the masking threshold value calculation unit 12 and the effective power spectrum extraction unit 13. Further, the time frequency conversion unit 11 outputs the frequency signal value of each frequency to the correction unit 16.

  The masking threshold value calculation unit 12 obtains a masking threshold value corresponding to a power spectrum value that is difficult for humans to perceive for each frequency (step S103). Then, the masking threshold value calculation unit 12 outputs the masking threshold value of each frequency to the effective power spectrum extraction unit 13. The effective power spectrum extraction unit 13 calculates an effective power spectrum value, which is a component that can be perceived by a person, by subtracting the masking threshold value from the power spectrum value for each frequency (step S104). The effective power spectrum extraction unit 13 outputs the effective power spectrum value of each frequency to the interband power difference calculation unit 14.

The inter-band power difference calculation unit 14 calculates a power difference ΔP between an average value of effective power spectrum values of each frequency in the low frequency band and an average value of effective power spectrum values of each frequency in the high frequency band (step) S105). Then, the interband power difference calculation unit 14 outputs the power difference ΔP to the correction amount calculation unit 15.
The correction amount calculation unit 15 determines a correction coefficient for each frequency in the correction target frequency band so that the frequency signal value of each frequency in the correction target frequency band is amplified more as the power difference ΔP is larger (step S106). ). Then, the correction unit 16 corrects the frequency signal value of each frequency by amplifying the frequency signal value for each frequency in the correction target frequency band according to the correction coefficient determined by the correction amount calculation unit 15 (step). S107). Then, the frequency time conversion unit 17 calculates the corrected audio signal by converting the corrected frequency signal value of each frequency into the time domain (step S108).
Then, the sound correction device 6 outputs the corrected sound signal and ends the sound correction process.

  As described above, this sound correction apparatus emphasizes the frequency signal value of each frequency in the high frequency band based on the difference between the components perceivable by humans in the power spectrum of the low frequency band and the high frequency band. Determine the degree. For this reason, the sound correction apparatus can appropriately improve the feeling of the sound signal bulk even if the difference between the power spectrum in the low frequency band and the power spectrum in the high frequency band is small.

  According to the modification, the masking threshold value calculation unit may calculate the masking threshold value only for the frequency included in the low frequency band and the frequency included in the high frequency band used for calculating the power difference. Similarly, the effective power spectrum extraction unit may calculate the effective power spectrum value only for the frequency included in the low frequency band and the frequency included in the high frequency band used for calculating the power difference. Thereby, the calculation amount is reduced.

  According to another modification, the sound correction apparatus may attenuate the frequency signal value of each frequency included in the low frequency band as the power difference between the low frequency band and the high frequency band is large. Also according to this modified example, the ratio of the frequency signal value of each frequency in the high frequency band to the frequency signal value of each frequency in the low frequency band becomes relatively high, so that the feeling of being muffled is improved. In the case of this modification, the correction amount calculation unit increases the attenuation coefficient for each frequency in the low frequency band as the power difference increases. And a correction | amendment part produces | generates a correction frequency signal by attenuating the frequency signal value of each frequency in a low frequency band, so that an attenuation coefficient is large.

  According to still another modification, the masking threshold value calculation unit may calculate the masking threshold value for each frequency using the absolute value of the amplitude of each frequency signal value instead of the power spectrum. The absolute value of the amplitude of the frequency signal value is also an example of a spectrum signal. In this case, the effective power spectrum extraction unit also obtains, as an effective spectrum signal, a value obtained by subtracting the masking threshold value from the absolute value of the amplitude of the frequency signal value for each frequency. The power difference calculation unit between the bands also calculates the average value of the effective spectrum signal of each frequency in the low frequency band and the effective spectrum signal of each frequency in the high frequency band, calculated based on the absolute value of the amplitude of the frequency signal value. Find the difference from the average value of. The correction amount calculation unit increases the correction coefficient for each frequency within the correction target frequency band as the difference increases.

  According to still another modification, the sound correction apparatus may improve sound distortion caused by a power spectrum in a high frequency band being too large. In this case, the correction amount calculation unit of the sound correction device performs correction so as to attenuate the spectrum signal value of each frequency included in the high frequency band when the power difference ΔP is smaller than the reference value P1. Determine the coefficient.

FIG. 10 is a diagram showing an example of the relationship between the power difference ΔP and the reference correction coefficient Gm according to this modification. In FIG. 10, the horizontal axis represents the power difference ΔP, and the vertical axis represents the reference correction coefficient Gm. A graph 1000 represents the relationship between the power difference ΔP and the reference correction coefficient Gm.
As shown in the graph 1000, in this modification, when the power difference ΔP is equal to or less than the reference value Pl, the reference correction coefficient Gm is set to a negative value, and the reference correction coefficient Gm decreases as ΔP decreases. When the power difference ΔP is equal to or less than the correction lower limit value Pmin, a negative constant value, for example, −10 dB is set. If the power difference ΔP is larger than the reference value Pl, the reference correction coefficient Gm is determined according to the power difference ΔP, as in the example shown in FIG.

  According to still another modification, the correction amount calculation unit may divide the high frequency band into a plurality of sub frequency bands and change the correction coefficient for each sub frequency band. For example, depending on the ability of a speaker, sound quality degradation such as sound cracking may occur when a spectrum signal in a high frequency band is excessively amplified. However, according to this modification, the audio correction apparatus can improve the volume in a range in which such sound quality deterioration does not occur.

なお、係数β1(f)及びβ2(f)は、それぞれ、低い方のサブ周波数帯域及び高い方のサブ周波数帯域に対応する。 For example, when the high frequency band is divided into two sub frequency bands, the correction coefficient g (f) is calculated according to the following equation.

The coefficients β1 (f) and β2 (f) correspond to the lower sub-frequency band and the higher sub-frequency band, respectively.

  FIG. 11 is a diagram showing the relationship between the frequency and the coefficients β1 (f) and β2 (f) shown in the equation (7). In FIG. 11, the horizontal axis represents the frequency, and the vertical axis represents the magnitudes of the coefficients β1 (f) and β2 (f). A graph 1100 represents the relationship between the frequency and the coefficient β1 (f). A graph 1101 represents the relationship between the frequency and the coefficient β2 (f). As shown in the graph 1100, the coefficient β1 (f) is 0 below the frequency Es1, and monotonously increases as the frequency increases above the frequency Es1 and below the frequency Eul1. The coefficient β1 (f) is constant above the frequency Eul1 and below the frequency Em. Furthermore, the coefficient β1 (f) is monotonously decreased as the frequency becomes higher than the frequency Em and below the frequency Ee. The coefficient β1 (f) becomes 0 when it becomes higher than the frequency Ee. On the other hand, as shown in the graph 1101, the coefficient β2 (f) is 0 below the frequency Es2, and monotonously increases as the frequency increases above the frequency Es2 and below the frequency Eul2. The coefficient β2 (f) becomes constant when it becomes higher than the frequency Eul2. However, Ee and Eul2 are substantially equal, and Es2 and Em are also set to substantially equal frequencies.

According to another modification, the sound correction device may perform sound correction processing on an input sound signal collected by a microphone mounted on a mobile phone and digitized by an analog / digital converter. Good. In this case, the input voice signal corrected by the voice correction device is output to the control unit of the mobile phone.
In this case, the lower limit ls and the upper limit le of the low frequency band may be set to the lower limit and the upper limit of the frequency band emphasized by the proximity effect of the microphone mounted on the mobile phone.
Furthermore, the audio correction device is not limited to a mobile phone, and may be mounted on a fixed phone, a telephone conference system, or the like.

  Furthermore, a computer program that causes a computer to realize each function of each unit of the sound correction apparatus according to each of the above embodiments is provided in a form recorded on a computer-readable medium such as a magnetic recording medium or an optical recording medium. Also good.

FIG. 12 is a configuration diagram of a computer that operates as a sound correction apparatus when a computer program that realizes the functions of the respective units of the sound correction apparatus according to the above-described embodiment or its modification is operated.
The computer 100 includes a user interface unit 101, a communication interface unit 102, a storage unit 103, a storage medium access device 104, and a processor 105. The processor 105 is connected to the user interface unit 101, the communication interface unit 102, the storage unit 103, and the storage medium access device 104 via, for example, a bus.

  The user interface unit 101 includes, for example, an input device such as a keyboard and a mouse, and a display device such as a liquid crystal display. Alternatively, the user interface unit 101 may include a device such as a touch panel display in which an input device and a display device are integrated. For example, the user interface unit 101 outputs an operation signal for starting the sound correction processing to the processor 105 in accordance with a user operation.

The communication interface unit 102 may include an audio interface for connecting the computer 100 to a microphone and a speaker and a control circuit thereof.
Furthermore, the communication interface unit 102 may include a communication interface for connecting to a communication network according to a communication standard such as Ethernet (registered trademark) and a control circuit thereof.
In this case, the communication interface unit 102 acquires an audio signal from another device connected to the communication network and passes it to the processor 105. Further, the communication interface unit 102 may output the corrected audio signal received from the processor 105 to another device via the communication network.

  The storage unit 103 includes, for example, a readable / writable semiconductor memory and a read-only semiconductor memory. And the memory | storage part 103 memorize | stores the computer program for performing the audio | voice correction | amendment process performed on the processor 105, and various data utilized by an audio | voice correction process.

  The storage medium access device 104 is a device that accesses a storage medium 106 such as a magnetic disk, a semiconductor memory card, and an optical storage medium. For example, the storage medium access device 104 reads a computer program for sound correction processing executed on the processor 105 stored in the storage medium 106 and passes the computer program to the processor 105.

  The processor 105 enhances the frequency components in the high frequency band so as to improve the feeling of the sound signal bulkiness by executing the sound correction processing computer program according to the above-described embodiment or modification. Then, the processor 105 outputs the corrected audio signal to other devices via the communication interface unit 102.

  All examples and specific terms listed herein are intended for instructional purposes to help the reader understand the concepts contributed by the inventor to the present invention and the promotion of the technology. It should be construed that it is not limited to the construction of any example herein, such specific examples and conditions, with respect to showing the superiority and inferiority of the present invention. Although embodiments of the present invention have been described in detail, it should be understood that various changes, substitutions and modifications can be made thereto without departing from the spirit and scope of the present invention.

The following supplementary notes are further disclosed regarding the embodiment described above and its modifications.
(Appendix 1)
A time-frequency conversion unit that calculates a spectrum signal including spectrum signal values for each of a plurality of frequencies by converting a time-domain audio signal into a frequency domain in units of frames having a predetermined time length;
For each frequency, a masking threshold value calculation unit that calculates a masking threshold value corresponding to a spectrum signal value that cannot be heard according to human auditory characteristics;
Perceptibility of the first frequency band and the second frequency band based on the spectral signal value and the masking threshold for at least the frequency included in the first frequency band and the frequency included in the second frequency band. An effective spectrum extraction unit for calculating an effective spectrum signal representing a simple spectrum signal;
An inter-band power difference calculation unit for obtaining a difference between the effective spectrum signal of the first frequency band and the effective spectrum signal of the second frequency band;
A correction amount calculation unit that determines a correction amount of a predetermined frequency band according to the difference;
A correction unit that calculates a corrected spectrum signal by correcting a spectrum signal value of each frequency in the predetermined frequency band according to the correction amount;
A frequency time conversion unit for obtaining a corrected audio signal by converting the corrected spectrum signal into the time domain;
An audio correction apparatus having
(Appendix 2)
The effective spectrum extraction unit subtracts the masking threshold value from the spectrum signal value for each frequency included in the first frequency band or the second frequency band, thereby the effective spectrum signal of the first frequency band. The speech correction apparatus according to appendix 1, wherein the effective spectrum of the second frequency band is calculated.
(Appendix 3)
The masking threshold calculation unit detects a peak frequency at which the spectrum signal value is a maximum value among the plurality of frequency bands in the current frame, and is close to the peak frequency and the spectrum signal value at the peak frequency is large. The audio correction apparatus according to appendix 1 or 2, wherein the masking threshold for the frequency is increased as the frequency is increased.
(Appendix 4)
The audio correction apparatus according to appendix 1 or 2, wherein the masking threshold value calculation unit increases the masking threshold value for the frequency as the frequency of the spectrum signal value in a frame a predetermined number before the current frame increases.
(Appendix 5)
The first frequency band is a frequency band in which the feeling of bulkiness is deteriorated by amplifying the spectrum signal value of the first frequency band, while the second frequency band is the second frequency band. The audio correction device according to any one of appendices 1 to 4, which is a frequency band in which a feeling of being obscured is improved by amplifying the spectrum signal value.
(Appendix 6)
The audio correction device according to any one of appendices 1 to 5, wherein the first frequency band is included in a third frequency band in which the frequency signal is amplified by characteristics of an audio input unit that has acquired the audio signal. .
(Appendix 7)
The second frequency band is higher than the first frequency band;
The correction unit increases the ratio of the spectrum signal value of the frequency in the second frequency band to the spectrum signal value of the frequency in the first frequency band according to the correction amount so as to increase the ratio. The audio correction device according to any one of appendices 1 to 6, which calculates a correction spectrum signal.
(Appendix 8)
The correction amount calculation unit determines the correction amount for at least one of the first frequency band and the second frequency band such that the greater the difference, the higher the ratio of the spectral signal values. The sound correction apparatus according to appendix 7.
(Appendix 9)
A spectrum signal including a spectrum signal value for each of a plurality of frequencies is calculated by converting a time domain audio signal into a frequency domain in units of frames having a predetermined time length,
For each frequency, calculate the masking threshold corresponding to the spectrum signal value that cannot be heard according to the human auditory characteristics,
Perceptibility of the first frequency band and the second frequency band based on the spectral signal value and the masking threshold for at least the frequency included in the first frequency band and the frequency included in the second frequency band. The effective spectrum signal representing the correct spectrum signal,
Determining a difference between the effective spectrum signal of the first frequency band and the effective spectrum signal of the second frequency band;
According to the difference, determine a correction amount for a predetermined frequency band,
According to the correction amount, a corrected spectrum signal is calculated by correcting a spectrum signal value of each frequency within the predetermined frequency band,
Obtaining a corrected audio signal by converting the corrected spectrum signal into the time domain;
An audio correction method including the above.
(Appendix 10)
A spectrum signal including a spectrum signal value for each of a plurality of frequencies is calculated by converting a time domain audio signal into a frequency domain in units of frames having a predetermined time length,
For each frequency, calculate the masking threshold corresponding to the spectrum signal value that cannot be heard according to the human auditory characteristics,
Perceptibility of the first frequency band and the second frequency band based on the spectral signal value and the masking threshold for at least the frequency included in the first frequency band and the frequency included in the second frequency band. The effective spectrum signal representing the correct spectrum signal,
Determining a difference between the effective spectrum signal of the first frequency band and the effective spectrum signal of the second frequency band;
According to the difference, determine a correction amount for a predetermined frequency band,
According to the correction amount, a corrected spectrum signal is calculated by correcting a spectrum signal value of each frequency within the predetermined frequency band,
Obtaining a corrected audio signal by converting the corrected spectrum signal into the time domain;
A computer program for sound correction that causes a computer to execute the operation.

DESCRIPTION OF SYMBOLS 1 Cellular phone 2 Control part 3 Communication part 4 Microphone 5 Analog / digital converter 6 Voice correction device 7 Digital / analog converter 8 Speaker 11 Time frequency conversion part 12 Masking threshold calculation part 13 Effective power spectrum extraction part 14 Interband power difference Calculation unit 15 Correction amount calculation unit 16 Correction unit 17 Frequency time conversion unit 100 Computer 101 User interface unit 102 Communication interface unit 103 Storage unit 104 Storage medium access device 105 Processor 106 Storage medium

Claims (8)

A time-frequency conversion unit that calculates a spectrum signal including spectrum signal values for each of a plurality of frequencies by converting a time-domain audio signal into a frequency domain in units of frames having a predetermined time length;
For each frequency, a masking threshold value calculation unit that calculates a masking threshold value corresponding to a spectrum signal value that cannot be heard according to human auditory characteristics;
Perceptibility of the first frequency band and the second frequency band based on the spectral signal value and the masking threshold for at least the frequency included in the first frequency band and the frequency included in the second frequency band. An effective spectrum extraction unit for calculating an effective spectrum signal representing a simple spectrum signal;
An inter-band power difference calculation unit for obtaining a difference between the effective spectrum signal of the first frequency band and the effective spectrum signal of the second frequency band;
A correction amount calculation unit that determines a correction amount of a predetermined frequency band according to the difference;
A correction unit that calculates a corrected spectrum signal by correcting a spectrum signal value of each frequency in the predetermined frequency band according to the correction amount;
A frequency time conversion unit for obtaining a corrected audio signal by converting the corrected spectrum signal into the time domain;
An audio correction apparatus having   The effective spectrum extraction unit subtracts the masking threshold value from the spectrum signal value for each frequency included in the first frequency band or the second frequency band, thereby the effective spectrum signal of the first frequency band. The sound correction apparatus according to claim 1, wherein the effective spectrum of the second frequency band is calculated.   The masking threshold calculation unit detects a peak frequency at which the spectrum signal value is a maximum value among the plurality of frequency bands in the current frame, and is close to the peak frequency and the spectrum signal value at the peak frequency is large. The sound correction apparatus according to claim 1, wherein the masking threshold for the frequency is increased as the frequency is increased.   3. The speech correction apparatus according to claim 1, wherein the masking threshold calculation unit increases the masking threshold for the frequency as the frequency of the spectrum signal value in a frame a predetermined number before the current frame increases. The second frequency band is higher than the first frequency band;
The correction unit increases the ratio of the spectrum signal value of the frequency in the second frequency band to the spectrum signal value of the frequency in the first frequency band according to the correction amount so as to increase the ratio. The sound correction apparatus according to claim 1, wherein the correction spectrum signal is calculated.   The correction amount calculation unit determines the correction amount for at least one of the first frequency band and the second frequency band such that the greater the difference, the higher the ratio of the spectral signal values. The sound correction apparatus according to claim 5. A spectrum signal including a spectrum signal value for each of a plurality of frequencies is calculated by converting a time domain audio signal into a frequency domain in units of frames having a predetermined time length,
For each frequency, calculate the masking threshold corresponding to the spectrum signal value that cannot be heard according to the human auditory characteristics,
Perceptibility of the first frequency band and the second frequency band based on the spectral signal value and the masking threshold for at least the frequency included in the first frequency band and the frequency included in the second frequency band. The effective spectrum signal representing the correct spectrum signal,
Determining a difference between the effective spectrum signal of the first frequency band and the effective spectrum signal of the second frequency band;
According to the difference, determine a correction amount for a predetermined frequency band,
According to the correction amount, a corrected spectrum signal is calculated by correcting a spectrum signal value of each frequency within the predetermined frequency band,
Obtaining a corrected audio signal by converting the corrected spectrum signal into the time domain;
An audio correction method including the above. A spectrum signal including a spectrum signal value for each of a plurality of frequencies is calculated by converting a time domain audio signal into a frequency domain in units of frames having a predetermined time length,
For each frequency, calculate the masking threshold corresponding to the spectrum signal value that cannot be heard according to the human auditory characteristics,
Perceptibility of the first frequency band and the second frequency band based on the spectral signal value and the masking threshold for at least the frequency included in the first frequency band and the frequency included in the second frequency band. The effective spectrum signal representing the correct spectrum signal,
Determining a difference between the effective spectrum signal of the first frequency band and the effective spectrum signal of the second frequency band;
According to the difference, determine a correction amount for a predetermined frequency band,
According to the correction amount, a corrected spectrum signal is calculated by correcting a spectrum signal value of each frequency within the predetermined frequency band,
Obtaining a corrected audio signal by converting the corrected spectrum signal into the time domain;
A computer program for sound correction that causes a computer to execute the operation.

Images